Why conviction matters in factor investing
A closer look at the role of conviction and why it is not fully captured by traditional approaches.
Key takeaways:
- Start with the portfolio, then infer the return beliefs needed to justify its risk trade-offs.
- Size alone misleads; adjust for volatility, correlation and constraints to reveal implied alphas.
- Convert implied beliefs into a consistent signal you can compare, combine and use in portfolio design.
In the previous article, we looked at why conviction matters to modern factor investing. If conviction is already embedded in portfolios, a natural question follows: can it actually be extracted and used more deliberately?
It turns out, it can. But not in the way you might expect.
Starting from the answer
Most investment processes follow a familiar path. You form expectations about returns, balance those expectations against risk, and then build a portfolio. The portfolio is the end result of that thinking.
This approach flips that process. Instead of starting with expectations and ending with a portfolio, you start with the portfolio and work backwards. That may sound unusual at first. We are used to thinking in a forward direction: forecast, optimise, implement. Here, the logic runs in reverse.
A portfolio is the outcome of a set of decisions but what is it that sits behind those decisions? At first glance, a portfolio looks like a set of holdings. But once you look more closely, it becomes something else – a set of trade-offs. Every position reflects a balance between expected return, risk and how that position interacts with others in the portfolio.
That is why simply reading position sizes at face value only tells part of the story. The visible structure is only the surface. The reasoning that created it sits underneath.
From decisions to expectations
To understand what a manager truly believes, you need to interpret those decisions in context. Position sizes are shaped by more than just return expectations. They reflect volatility (how much a stock tends to move), correlation (how it moves relative to other holdings), and the role each position plays within the broader portfolio.
This is where most approaches stop. They observe the positions and infer conviction directly from size, but our approach goes further.
Once you account for those factors, a clearer picture starts to emerge. You can begin to identify which ideas are doing the real work, where risk is being taken deliberately, and where conviction is strongest. At that point, something important happens and you move from observing positions to uncovering expectations. Not what is held but what is expected.
Those expectations are not directly visible. But they are embedded in how the portfolio has been constructed. With a risk-aware framework (one that interprets positions in light of volatility, correlation and constraints) it becomes possible to infer those expectations in a consistent way.
These inferred expectations are often referred to as ‘implied alphas’. In simple terms, a set of return expectations that would make the observed portfolio make sense given its risks and trade-offs.
Turning insight into something usable
Once those embedded expectations are uncovered, they can be translated into a consistent signal. A signal that reflects how strongly different ideas are backed, adjusted for the role they play in the portfolio.
Instead of asking a manager what they expect, you are inferring what they must believe based on how they have allocated capital.
This matters because it makes conviction usable.
Instead of being locked inside individual portfolios, those insights can now be compared across managers, combined into a single view and applied more deliberately within a broader investment process.
That shift is subtle, but powerful. The insight itself does not change. What changes is the form it takes and what you can do with it.
Rather than working with entire portfolios as fixed building blocks, you are now working with the underlying expectations that drive them. That makes it possible to separate stock-level insight from the structure of the original portfolio.
What this changes in practice
Once conviction can be captured in this way, portfolio construction starts to look different.
Rather than combining managers and inheriting the structure of each portfolio, you can focus on the underlying ideas. You can bring together high-conviction views from different sources, while still maintaining control over risk, diversification and alignment to a benchmark.
This helps address a familiar problem. Instead of diluting conviction as portfolios are combined, you can preserve the strength of individual insights while still building a coherent overall portfolio.
It also creates more flexibility. Portfolios can be adjusted, refined and tailored to specific objectives without losing the underlying signal that made those ideas valuable in the first place.
The result is a portfolio that is both structured and expressive.
A portfolio that retains the discipline of systematic investing while making better use of the strongest insights.
What comes next?
The mechanics behind this approach are more nuanced than they first appear but the core idea is simple.
The insight you are looking for is already there,
the question is whether you are using it.
In the full Capturing Conviction paper, we explore how this approach works in detail and how those signals can be combined, refined and applied.
